we recognize) this condition is unlikely to hold, so that our conception of collab-
orative governance is chieXy a phenomenon of relatively healthy polities. Second,
each of the collaborating parties must possess a degree of discretion. If private
participants merely carry out government’s instructions—conveyed through fully
speciWed contracts or other means—the relationship is something other than col-
laborative governance.
Indeed, the allocation of discretion is the most useful discriminant for separating
collaborative governance from other forms of public–private interaction. Consider,
on the one hand, corporate charitable contributions. Companies enjoy broad dis-
cretion over their philanthropic giving, and their choices are presumptively deWned
as ‘‘the public good’’ for tax purposes. There are limits, to be sure. Charitable
deductions cannot, under current law, exceed 10 per cent of taxable corporate income
(a constraint that rarely binds). No deductions can be claimed for gifts to political
parties, or to the CEO’s shiftless cousin. But while shareholders might quibble over
grants to the chairman’s alma mater, or the local polo league, or exotic religious sects,
the government has no standing to complain short of trying to discredit the charity
itself. The public sector is a party to the undertaking—surrendering revenue it would
have otherwise received—but is a passive and silent partner. No doubt this arrange-
ment permits occasions of waste or triviality, but there are strong reasons for
protecting donors’ discretion against governmental second-guessing on the merits
of the mission—for example, so that government does notWnd itself in the position
of declaring which religions are acceptable and which are not. (The Comptroller of
Texas attempted to strip Unitarianism—one of America’s oldest denominations–of
its status as a tax-exempt church in 2004 , on the grounds of excessive heterodoxy, but
reconsidered after mild local protests and louder national ridicule (Herman 2004 ).)
Consider, conversely, a municipal government contracting with a private waste
management company. The company’s mission—to pick up the garbage and dump it
at the landWll—is explicit, complete, and controlled by the government, and its
motive is to maximize the revenue (less costs) it receives in return. If, upon contract
renewal, the government wants the garbage to be collected on Fridays instead of
Wednesdays, or incinerated instead of buried, it is at liberty to alter the mandate and
the company’s only legitimate claim is fair payment for the work. The private player
is a pure agent, and discretion rests with the government. Denying the agent
initiative—e.g. the right to shop for the cheapest disposal option—obviates some
of the reasons for engaging private agents in theWrst place. But in many arenas of
public–private interaction such one-sided discretion is both customary and prudent.
We do not address the myriad complexities that attend pure voluntarism or pure
contracting. Nor do we suggest that binary assignments of discretion—wholly
private or wholly public—are the normal case. Our goal is to demarcate the domain
within which collaborative governance resides, and to underscore that the sharing of
discretion both enriches the potential of public–private interaction and renders it
much more complex, not just in application but analytically, in ways we will seek to
describe once a few examples introduce somewhat more concreteness into the
discussion.
510 john d. donahue & richard j. zeckhauser